csaf financial benchmarking - agrilinks · the csaf financial benchmarking analysis suggests the...
Post on 02-Jun-2020
4 Views
Preview:
TRANSCRIPT
CSAF financial benchmarkingFinal learning report
July 2018
Executive Summary (1/4): Context
• An estimated 75% of the market demand for smallholder agricultural finance is not being served, meeting only an estimated $50 billion of the more than
$200 billion demand across Latin America, Sub-Saharan Africa, and South and Southeast Asia.1 This market sizing includes financing needs both at the farmer
level and for SMEs operating in agricultural value chains. Agricultural small and medium-sized enterprises (SMEs) provide important services such as access
to inputs, credit, agronomic training, and market linkages for smallholder farmers and also create employment via value-added processing, transportation, and
other services. Access to finance for smallholder farmers and SMEs along agricultural value chains is critical to boost agricultural productivity, improve farmer
livelihoods, create jobs, strengthen food security & nutrition, and build resilience to climate change.2
• Given the characteristics of rural areas and agriculture-based economic activities, lending to agricultural SMEs faces a number of challenges, in
addition to those inherent in any financial intermediation. These specific challenges are related to crop seasonality and price volatility, external risks like
climate change or crop disease, and higher cost to serve low population densities. In less developed markets with poor physical infrastructure (e.g., sub-par
roads, electricity, ports) and a weak enabling environment (e.g., non-existent or inadequate insurance markets, disaster relief, collateral registry, credit bureaus,
arbitration), lenders serving agricultural SMEs feel these challenges even more acutely.
• Preliminary evaluations indicate that extending finance to credit-constrained agricultural SMEs generates significant impact for both enterprises
and farmers. An academic evaluation conducted on lending by one member of the Council on Smallholder Agricultural Finance (CSAF) found that access to
finance for credit-constrained enterprises directly increases enterprise growth as SMEs purchase more volume from farmers and at higher prices than they
would have otherwise.3 In a separate analysis of its own data, one CSAF lender found that half of enterprises borrowed for five consecutive years and those
borrowers increased their sales at an annual rate of 24 percent over this period. The same lender has conducted impact studies on 30 borrower enterprises
and found that incremental income to farmers affiliated with these enterprises typically ranges from 10-25 percent, with some variation on either end.
• Knowledge about capital supply and demand in the agricultural sector has primarily been concentrated in the direct-to-farmer segment of the
market. Much less is known about the market size or segmentation of agricultural SMEs, or the financial performance of intermediaries serving agricultural
SMEs in the “missing middle” – the gap between microfinance and lending by commercial banks to larger enterprises.
• Based on empirical observations that the missing middle persists in the agriculture sector and anecdotal evidence suggesting that the economics of
those few lenders serving the segment are challenging at best, we hypothesize that public and philanthropic interventions are required to unlock market
development. Donors have invested in building markets like microfinance. The question is whether and how to do the same for agricultural SME finance. Given
the current state of information about the agricultural SME finance market, it is difficult to identify where interventions are most urgently required, design
solutions, and mobilize stakeholders to fund and implement them.
2
(1) Initiative for Smallholder Farmer Finance and Rural Agriculture Finance Learning Lab, “Inflection Point: Unlocking
growth in the era of farmer finance”, 2015
(2) Further supporting evidence in appendix, page 64
(3) Blouin, Arthur and Macchiavello, Rocco, “Strategic Default in the International Coffee Market," Working Paper F-
89305-CCP-1, International Growth Centre, 2017.
Executive Summary (2/4): Project approach and findings
(1) Initiative for Smallholder Farmer Finance and Rural Agriculture Finance Learning Lab, “Inflection Point: Unlocking
growth in the era of farmer finance”, 2015
• To inform this dialogue, data about the costs and risk to serve agricultural SMEs, as well as the impact generated when they have reliable access to
finance are required. However, few quantitative assessments of the financial performance of lenders serving agricultural SMEs are available. Obtaining this
data from lenders is difficult due to data security concerns, competition, and complexity of standardizing and analyzing the data. The analysis shared in this
report draws upon an original and rich dataset and represents a leap forward in understanding the economics of the nascent agricultural SME finance industry.
• To perform this first-of-its-kind analysis, and quantify the economics of lending to agricultural SMEs, Dalberg worked with nine members of the
Council on Smallholder Agriculture Finance (CSAF) to gather information and create a unique database of lending performance by segment. The
database brings together information on 3,556 individual loans, including transaction revenues and costs, operating costs, and impact-adjusted costs of funds,
allowing the analysis of the conditions and segments for profitability and unprofitability of lending in this market. This report was the first phase of work focused
on social lending organizations which are members of CSAF. A second phase from June through August will survey local financial institutions as well, to
provide data that is representative of the local market and give a more comprehensive view on lending economics of commercial lenders in sub-Saharan Africa.
• The economics of lending to agricultural SMEs are challenging.
On average, a CSAF loan analyzed lost ~$1,000, representing 0.34
percent of the average loan size of $665,000.
– The average income from loans analyzed was ~$42,900,
representing an annualized interest + fee yield p.a. of 8.1 percent.
– The average operating cost per loan was ~$23,800, including costs
of origination, servicing and allocated overheads.
– The average credit losses including recovery costs was ~$20,000.
The majority of these costs came from the <10 percent of the loans
with associated recoveries.
– The average cost of funds associated with the loans was $16,100
assuming an average 3% p.a. below-market capital for social lending
• Nevertheless, over 50 percent of the CSAF loans analyzed were
profitable (before accounting for costs of funds). These loans tended
to cluster in specific segments, such as loan sizes larger than
$500,000, and loans in coffee and cocoa, which were were profitable
on average for CSAF lenders.
$2.6k
Income net
of op. costs
-$23.8k
$18.2k
Loan
transaction
revenue
$38.6k
$4.2k
Operating
costs
Credit
losses and
recovery
costs
Income net
of credit
losses
$16.1k
Risk-
adjusted
impact cost
of funds
Income
net of cost
of funds
$19.0k
-$1.8k
-$17.9k
Average
annualized
fee yield
p.a.: 0.8%
Average annualized interest yield p.a.:
7.3% (after currency losses of 0.4%)
Loan economics averages for all CSAF loans analyzed
USD thousandsCurrency losses=
Average
annualized
write-off
3.9%
3
Executive Summary (3/4): Segment findings
4
• The analysis revealed large variations in financial performance across the different segments of the agricultural SME market. Using five testable
segmentation drivers, we traced the differences in loan-level profitability by region, facility size, borrower status, value chain, and loan tenor. The findings are
summarized below.
– Region: On average, CSAF loans in Sub-Saharan Africa were less profitable than loans in other regions. Loans in Sub-Saharan Africa represent only
one-fifth of overall loans analyzed and show lower profitability due to higher credit losses. They were also twice as likely to end up in recovery, and had 22
percent higher operating costs than in other geographies, resulting in net losses (after credit losses and overheads) for the loans analyzed.
– Facility size: Of the CSAF loans analyzed, smaller loans tended to have lower profitability. While over half the loans disbursed were below $500,000,
they showed lower profitability. This was driven by the lower interest income, even though operating costs incurred were the same. They also showed an 80
percent higher risk of impairment than those above $500,000. Given these factors, the loans under $500,000 (after credit losses and overheads) had net
losses.
– First-time borrowers: Loans to new borrowers were significantly less profitable than those to existing borrowers. Less than one-quarter of the
loans analyzed went to new borrowers; these were loss making on average due to higher risk and costs to serve. The data and surveys from the lenders
showed that new borrowers incurred 50 percent higher origination costs and had twice the risk of impairment. Consequently, first-time borrower lending
(after credit losses and overheads) had net losses on average.
– Value chain: Lending in coffee and cocoa value chains was more profitable than lending in other crops. Two-thirds of all capital lent over the period
analyzed went to the coffee or cocoa value chains, perhaps because of their more mature and predictable market. Analysis showed that loans outside the
coffee and cocoa value chains carried 2.5 times the risk of impairment, and several lenders also reported higher origination costs for them. As a result, loans
to crops other than coffee and cocoa were marginally loss-making, while coffee and cocoa loans (after credit losses and overheads) yield a positive return.
– Tenor: Long-term loans (12 months or more) were less profitable than shorter-term lending (less than 12 months). Nearly three-quarters of loans
analyzed were lent for short-term financial needs. Given the unpredictability and irregularity of cash flows in agriculture, longer-term lending may be
perceived as higher risk. Of those analyzed, loans with tenors of more than 12 months were unprofitable on average and had a 4 times higher risk of
impairment. Long-term loans incurred net losses on average (after credit losses and overheads)
• Lower profitability in some segments versus others has three primary causes: lower revenues due to lower interest or fee revenue, higher operating
costs to serve, and/or higher risks due to increased likelihood of delinquency or impairment.
(1) Detailed breakdown of profitability analyses in appendix tables, pages 74-75
Executive Summary (4/4): Implications
• Evidence suggests improved capital flows to agricultural SMEs can help boost smallholder farmer productivity, incomes and resiliency to shocks1.
The CSAF financial benchmarking analysis suggests the economics of providing capital is hard, particularly in some segments. Donors could play a critical role
to help bridge the financing gap for agricultural SMEs serving smallholder farmers by supporting initiatives that can improve the sustainability of lending in high-
impact segments through blended finance tools and other supporting mechanisms to close the financing gap.
• Donor interventions to stimulate the agricultural finance market should adhere to certain principles: 1) Market-orientation: to incentivize competition,
efficiency, and innovation that will drive down the requirement for subsidy over time, 2) Additionality: to complement current market activity, and maximize
participation of private capital, and 3) Alignment with impact objectives. Four types of blended finance instruments that can unlock the flow of finance to
agricultural SMEs serving smallholder farmers can be explored:
– Output-based incentives: A financial incentive facility can be used to promote financial access to segments with low or negative profitability but high
impact potential. Such incentives could follow a pay-per-loan model based on tiered scoring of loan characteristics and the likelihood of reaching certain
impact goals, such as lending to smaller SMEs (with smaller financing needs).
– Risk mitigation: This financial incentive allows lenders to explore riskier segments. A facility such as partial credit guarantees and/or “first-loss” buffers,
which might be appropriate for different segments, may be considered to encourage lending in new, underserved segments perceived to be high risk, such
as first-time borrowers or long-term lending.
– Direct funding: Providing balance sheet support or concessional funding for lenders to increase their risk appetite frees up capital for lending to high-impact
segments with higher perceived and real risk. Such segments could include loose value chains or frontier markets.
– Technical assistance: In addition to direct financial support to lenders, offering advisory support to lenders can help lower their operating costs, and to
borrowers to can help reduce their risk profile, especially in high costs segments, such as first-time borrowers and SMEs in loose value chains.
• A broader set of donor actions taken in concert could improve the attractiveness of agricultural SME loans. Other initiatives might include providing
funding for disruptive technological innovations and/or promoting competition from new actors with potentially disruptive business models. Finally, donors could
focus on providing highly coordinated value chain or enabling environment interventions to help lower transaction costs, increase scale and/or reduce risk.
• By the end of Q3 2018, we expect to have a broader dataset that includes local banks and non-bank financial institutions in East Africa. This data
provides, for the first time, transparency on the current financial performance of lenders serving agricultural SMEs and allows for frank discussions about what
interventions might be required to catalyze a more competitive and sustainable market.
5
(1) International Finance Corporation, Innovative Agricultural SME Finance Models, (2012), and Rural Agricultural
Finance Learning Lab, Evidence on the Impact of Rural and Agricultural Finance on Clients in Sub-Saharan Africa: a
Literature Review (2015).
Agenda
6
1. Introduction
2. Approach
3. Key findings
4. Implications
5. Appendix
Financial needs and disbursements of commercial smallholder*
farmers by value chain (USD Billion)1
Three quarters of smallholder farmer financial needs remain unaddressed
due to market and sector characteristics
Several agriculture and SME-related risks and costs deter lenders
from bridging this financing gap2:
• High risk: seasonality and uncertainties of crop production, currency
volatility implying high hedging costs, crop price volatility, climate change
and increased extreme weather events
• High servicing costs: providing financial services to rural populations
can lead to higher transaction costs, and FSPs often lack knowledge on
how to manage these costs
24%12%
76%88%
~$65B ~$90B
Tight
(~18M farmers)
Loose
(~88M farmers)
Available financing
Unmet demand
The above factors factors weigh more heavily when lending to the
most vulnerable segments of the market:
• More risk in loose value chains as opposed to tight value chains3
i.e., formalized, consolidated markets with clear standards and specific
contractual obligations (e.g., coffee)
• More risk for new borrowers due to lack of track record
• More risk for long term loans due to the volatility of agriculture
markets4
7
(1) Inflection Point: Unlocking growth in the era of farmer finance, Initiative for Smallholder Finance and Rural Agriculture
Finance Learning Lab and executed by Dalberg, 2016 (2) Risk in Agriculture, USDA, 2015; World Bank, Agriculture Finance,
2015 (3) Segmentation of Smallholder Households: Meeting the Range of Financial Needs in Agricultural Families, CGAP,
2013 (4) Catalyzing Smallholder Agricultural Finance, Dalberg, 2012
Notes: (*) Supply includes formal, informal fin. institutions and value chain actors, demand includes agricultural and non
agricultural needs of smallholder farmers
$210B
$15B
$25B
$78B
$2B
$63B
$27B
Total smallholder
finance need
$54B
Unmet demandAvailable
financing
$156BNon-agri needs
Short-term needs
Long-term needs
Demand for smallholder farmer financing
(Annual disbursements, USD billion)1
Over 75% of the
financing demand
remains unmet by both
formal and informal
actors
Smallholder farmer finance meets only around $50 billion of the $200
billion demand across Latin America, Sub-Saharan Africa, &
South/Southeast Asia*. There is also an important, not yet sized need at the
agricultural SME level, according to CSAF members.
Financial need is greatest for the “missing middle”1, i.e., SMEs that serve
smallholder farmers with capital needs between $50K and $1M USD
8
Commercial banks:
• Typically lend from $1M and above
• Usually require fixed asset collateral
(1)The Elephant in the Room: Financial Inclusion for the Missing Middle, 2015
(2) Graphic courtesy of CSAF
(3) Initiative for Smallholder Finance, “A Roadmap For Growth: Positioning Local Banks For Success In Smallholder
Finance,” 2013
(4) Illustrative diagram courtesy of CSAF
$10k
Existing Market
$5M
$2M
$500k
$100k
Risk Factors Frontier Markets
Social lenders
Commercial Bank Lending
“Missing
middle”
Microfinance Institution Lending
Microfinance Institutions3:
• Lend at a very small ticket size
• Moving towards higher loan sizes while
remaining well under $50k
Social lenders:
• Lending from $100K -$2M,
• Extending beyond commercial banks to
reach a portion of the missing middle
• Often provide unsecured lending tied to
seasonal production in absence of formal
collateral
Illustrative representation of the state of the market in 20182
Loan size, USD
This illustrative representation only refers to agricultural SMEs. An important financing gap also exists in
direct financing for individual smallholder farmers.
ILLUSTRATIVE EXAMPLE4
This study helps bridge the knowledge gap on the financial performance of
actors lending at the frontier of the “missing middle”
9
• While research is available on smallholder farmers financial needs, studies of agricultural SME demand are insufficient. The five
categories of lenders that address some of the smallholder farmers’ needs are:
– Value chain actors that provide in-kind services or short-term cash advances based on their relationships with farmers
– Microfinance institutions that traditionally provide non-agri group lending and savings (and have recently moved into short-term agriculture lending)
– State banks that provide a high proportion of savings accounts and short term lending, but have very different funding and cost structures from
those of private entities, and have very limited presence (if any) in Africa
– Commercial banks that provide very small loans at high interest rates as well as larger loans requiring fixed asset guarantees; they rarely serve the
most vulnerable parts of the markets
– Impact investors and non-bank financial institutions, including social lenders and local finance companies specializing in agri-SMEs, providing
tailored financial services to an important array of underserved markets segments
• Most existing studies do not focus on the specific challenges of lenders (social or otherwise) covering the “missing middle”, and
quantitative assessments of lending performance are even fewer. Getting an in-depth understanding of the true challenges of lending to
agricultural SMEs with capital needs between $50K and $1M requires detailed performance data, which is limited given its sensitive nature, and
the small pool of actors. A first step towards uncovering the financing gap, was taken in 2016 with Initiative for Smallholder Finance report,
“Inflection Point: Unlocking growth in the era of farmer finance”. This reports builds on it further, utilizing the same approach to segmentation,
choice of assumptions, and recommendations for this financial benchmarking exercise.
• For this report, Dalberg built a unique, first-of-its-kind, database of loan-level profitability data to assess the ‘true cost’ of lending to
agri-SMEs within different segments. From the analyses, Dalberg assessed the relative risk associated with each segment, and illustrated
potential funding mechanisms for donors to help address the gaps. The results refine previous analyses and confirm some of the broader
tendencies observed in the market.
9
This engagement helps to better understand actual financial performance of CSAF organizations, whether there
is a need for a blended finance facility and in what situations such a facility might be deployed most useful.
Source: Initiative for Smallholder Finance, RAF Learning Lab, and Dalberg, “Inflection Point: Unlocking growth in the era
of farmer finance,” 2016; Dalberg ,“Catalyzing Smallholder Agricultural Finance”, 2012; FAO ESA Working Paper No. 14-
02; expert interviews; Dalberg analysis, CCAFS , “Scaling up index insurance for smallholder farmers”, CARE, Plan,
Barclays, “The State of Linkage Report: the first global mapping of savings group linkage”
Agenda
10
1. Introduction
2. Approach
3. Key findings
4. Implications
5. Appendix
Dalberg designed a two-phase study to determine the ‘true cost’ of SME
lending in agriculture and analyze its profitability drivers
11
Phase 1 Phase 2
A financial benchmarking analysis of the portfolios of
the Council of Smallholder Agricultural Financing
(CSAF)1 lenders to assess the cost to serve different
SME segments. We analyzed:
• Revenue of loans, in terms of interest income and
fees
• Financial risk of write-offs
• Operating costs, including one-time and ongoing,
direct and indirect costs
Based on the findings of Phase 1, funded by USAID,
we identified drivers of margins, cost, and risk, as well
as potential opportunities for donor interventions
Phase 2 will replicate the Phase 1 financial benchmarking
analysis for 8-10 local lenders, including banks and non-
bank financial institutions in East Africa to develop a
comprehensive view on lending economics and compare
with CSAF findings. Additional analyses on the research
agenda also include:
• Customer lifetime value in agri-lending: enterprise
growth and lender economics of long-term lending
relationships
• Technical assistance: costs and benefits of technical
assistance to agri-SME lending
• CSAF additionality analysis: mapping CSAF lending to
local lenders
(1) CSAF is a pre-competitive alliance of 12 financial institutions that promote industry standards and best practices for a
thriving and sustainable market serving the financing needs of agricultural SMEs globally. Members include AgDevCo,
Alterfin, Global Partnerships, Impact Finance, Incofin Investment Management, Oikocredit, Rabobank’s Rabo Rural
Fund, responsAbility Investments AG, Root Capital, Shared Interest Society, SME Impact Fund, and Triodos Investment
Management.
Dalberg analyzed the financial data of ~3.6k loan transactions from nine CSAF
lenders in Phase 1 to benchmark performance in agricultural SME lending
12
• Dalberg surveyed nine CSAF lenders to
gather the following data on their
agriculture lending portfolio from 2010-
2017 in three areas:
– Loan-level time series data: schedule
of loan disbursements and
repayments, including fees, interest,
and credit losses
– Portfolio breakdown of loan
characteristics: borrower details such
as country, value chain, facility type,
etc.
– Operating cost data: annual cost data
by region / business unit where
possible, including compensation,
legal and professional fees, back-
office resources, and other
overheads
Collect data Standardize Analyze
• Dalberg cleansed the loan data to arrive
at 3,556 in-scope loans, categorizing
value chains, facility types, etc.
• A weighting factor (dollar duration) was
utilized to allow a like-for-like
comparisons of profitability drivers
• The total annual operating costs were
divided across the originated and active
portfolio for each year, and allocated
across the stages of the loan lifecycle
• Dalberg validated initial loan analyses
as well as cost allocations with each
lender through bilateral conversations,
surveys, and other validation exercises
• Using the cleaned, standardized data,
Dalberg determined the financial profit
and accounting profit for each of the
loans provided by the CSAF lenders
• Dalberg also calculated the cost of
funds using for an impact-oriented
lender at an average of 3% p.a. (based
on discussion with CSAF lenders) to
determine the income net of an impact-
oriented cost of funds
• This resulted in unique and anonymized
database that allowed analyses of the
lending economics for serving
agriculture SMEs across the variety of
parameters and segments collected on
the portfolio
Further details of the methodology can be found in the appendix, page 67-73
The resulting data was broken down to determine net-profit for an average
loan in the CSAF dataset, and later analyzed within different segments
13
Loan economics averages for all CSAF loans analyzed1
USD thousandsAverage loan size: ~$665,000
Loan transaction
data
Op. cost
data
Loan transaction
& Op. cost data Dalberg analysis
So
urc
e
Income net
of operating
costs
$42.8k
Loan transaction
revenue
(fees + interest)
$20.7k
Income net
of credit losses
$23.8k
Operating costs
$19.0k
Credit losses and
recovery costs
$1.8k
$16.1k
Risk-adjusted
impact cost
of funds
$17.9k
Income net
of cost
of funds
Origination costs
+Servicing costs
+Allocated fixed
costs
Currency loss
Further details
on the following
page
(1) Calculated based on averaging each individual metric across all loans in a given dataset; all analysis with this title
utilize the same methodology (but with potentially different datasets depending on segmentation)
Based on this breakdown we have analyzed drivers of profitability variation across different segments
We use an 'impact cost of funds' based on the below-market capital that
the CSAF lenders raise to serve frontier markets
14
In order to account for the cost borne by a lender to raise capital, the
profitability analysis required an assumption on the average cost of funds.
There were thee potential scenarios by which Dalberg considered to
determine the cost of funds given the nature of the lenders and profile of
the borrowers (the impact of each illustrated on the right):
• Ignore cost of funds altogether and focus on operational profits
• Assume an ‘impact” cost of funds (~3% p.a.) to reflect the below-
market capital that social lenders, such as the CSAF lenders, would
access from impact investors
• Assume a commercial cost of funds (~6% p.a.) to reflect the costs
any lender operating within commercial markets would incur
Dalberg chose to focus on impact cost of funds, and took two main steps to
estimate it:
• Dalberg set average cost of funds at 3% p.a. based on discussions
with CSAF lenders
• This average was risk adjusted on a loan-by-loan basis based on the
profile of each loan (e.g., region) and sample averages
1
2
3
Used for analysis in this report
$16.1k
Loa
n tra
nsa
ction
re
ve
nu
e
$45.4k
Op
era
ting
co
sts
and
cre
dit lo
sses
Ad
ditio
na
l co
mm
erc
ial C
OF
Incom
e n
et o
f co
sts
and
write
-off
s
Ris
k a
dju
ste
d Im
pact C
OF
Incom
e n
et o
f Im
pact C
OF
$17.0k
Incom
e n
et o
f co
mm
erc
ial C
OF
-$35.7k
-$18.7k
$42.8k
-$2.6k
Loan economics averages for all CSAF loans analyzed
USD thousands
Average loan size: ~$665,000
1 2 3
For purposes of this report, we have defined simplified segments and used
other standardization metrics to best illustrate the underlying patterns
15
Small vs. large
loan sizes
S/Saharan
Africa vs. other
regions
New vs.
existing
borrower
Loose vs. tight
value chains1
Long-term vs.
short-term
Segment definitions
• Small loans: Loan sizes of less than $500,000
• Large loans: Loan sizes of greater than
$500,000
• S/Saharan Africa: Loans disbursed to
borrowers in sub-Saharan African countries
• Other regions: Loans disbursed in all other
countries in Latin America and Asia
• New borrower: Loans disbursed to a borrower
borrowing from the lender for the first-time
• Existing borrower: Loans disbursed to
borrowers who have borrowed from the lender
before
• Loose value chains: Loans in value chains
other than coffee and cocoa
• Tight value chains: Loans in coffee or cocoa
value chains
• Long-term loans: Loans with tenors greater
than 12 months
• Short-term loans: Loans with tenors lesser than
12 months
Duration
(months)
Dollar-duration
/ Weighting
factor ($)
Annualized
yield p.a. (%
per $ per year)
Additional analysis metrics2
(1) Segment definition adapted from Innovative Agricultural SME Finance Models (Global Partnership
for Financial Inclusion, International Finance Corporation)
(2) Further details on calculations and concepts in appendix, page 73 and 77
• Average number of months that a given dollar of
principal is outstanding
• For example, $1M loan being repaid in $500k
increments after 6 and 12 months has duration
of 9 months
• Product of the duration (defined above) of the
loan and the total amount disbursed
• For example, any loan with a $1 dollar-duration
is equivalent to a loan of $1 that is fully
outstanding for exactly one year
• The total amount of income as a proportion of
the total dollar-duration of the portfolio. Income
may be fees, interest, profit, or credit losses.
• For example, fee income yield p.a. of 1%
means that for every dollar that stays
outstanding for a year, 1 cents will be received
in fee income.
Agenda
16
1. Introduction
2. Approach
3. Key findings
• Sub-Saharan Africa vs. Rest of the World
• Small vs. large loan sizes
• New vs. existing borrowers
• Loose vs. tight value chains
• Long-term vs. short-term loans
4. Implications
5. Appendix
The largest share of loans analyzed were for working capital in Latin America
in the coffee value chain to existing borrowers
Notes: (*) borrowers that have previously accessed a loan from the same lender
(1) Further detail in appendix, page 74
26%
46%
17%
12%
Working Capital 6-12 months
Working Capital <6 months
Working Capital >12 months
Asset finance equipment
29%
30%25%
17%
>$1M
<$250k
$250-500k
$500k-1M
24% 76%
Breakdown of all loans disbursed 2010-2016
Percentage of loans, by region, loan size, value chain, financing product and new vs. existing borrower1
New borrower Existing borrower*
17
57%8%
26%
3%
2%
2%2%
Coffee
Cashew nuts
Cocoa
Quinoa
Honey
Sesame
Other crops
69%
23%
8%
Asia
Latin America/Caribbean
Africa
The agriculture portfolio analyzed was clustered in certain segments
81%
19%
Rest of the World
S/Saharan Africa
76%
≤$500k
>$500k
24%
Existing
80%
New
20%
66%
Other
34%
Coffee / Cocoa
73%
≤12 months
27%
>12 months
Loans in Sub-
Saharan Africa
represented less
than one fifth of
capital lent out
Loans smaller than
$500k in size
represented only one
quarter of capital
disbursed
Only 20% of capital
was utilized to lend
to a first-time
borrower
Nearly two-thirds of
capital was lent in
tight value chains,
i.e., coffee and
cocoa
Long-term lending
(>12 months) made
up only 27% of all
funds provided
The SME financing
demand in Africa
may continue to
remain underserved
Smaller SMEs with
smaller capital needs
may continue to
have their financing
needs go unmet
SMEs without prior
credit history may
continue to remain
out without access to
lending
SMEs in loose value
chains, which may
have greater needs,
could be
underserved
Long-term
investments required
to spur growth of
SMEs may
potentially be
underserved
% of funds
provided
over period
Implication
18Source: CSAF lenders survey conducted between April – June, 2018 of 3,556 individual loan transactions
(1) For purposes of this report, loans in coffee and cocoa value chains are considered as ‘tight’ while all others were
considered as ‘loose’
Small vs. large
loan sizes
S/Saharan Africa
vs. Rest of World
New vs. existing
borrower
Loose vs. tight
value chains1
Long-term vs.
short-term loans
Each of the segments was found to correlate with higher profitability
19
Net profit1
(USD,
thousands)
(1) Net profit = Interest + Fees – credit losses – Operating costs – Recovery costs – Currency losses;
Excludes cost of funds
(2) Annualized figures weighted on dollar-duration
Oth
er
$21.0k
New
borr
ow
er
Rest of th
e W
orld
Sub-S
ahara
n A
fric
a
>$500k
-$23.9k
Coff
ee/C
ocoa
≤12 m
onth
s
≤$500k
Exis
ting b
orr
ow
er
>12 m
onth
s
$3.6k
-$20.7k-$17.9k
$5.4k
$0.6k
-$6.2k
$1.9k
-$12.0k
0.7% -4.6% 2.2% -8.2% 1.1% -3.6% 0.2% -1.1% 0.6% -1.1%Annualized
net profit1
(% p.a.)2
Small vs. large
loan sizes
S/Saharan Africa
vs. Rest of World
New vs. existing
borrower
Loose vs. tight
value chains
Long-term vs.
short-term loans
Larger loan sizes tend to be more profitable across CSAF, and more than
half of loans <$500K would be loss-making even with zero-cost capital
20
-100
10
-10,000
100 10,0001,000
-1,000
-10
-1
0
1
10
100
1,000
10,000
,
Profitable
Non-profitable
Profitability
Net profit1; percentage of loans in segment that are profitable (excluding cost of funds), by loan size (USD thousands, log scale)
(1) Net profit = Interest + Fees – credit losses – Operating costs – Recovery costs – Currency losses;
Excludes cost of funds
Source: CSAF lenders survey conducted between April – June, 2018 of 3,556 individual loan transactions
Loan size ($k)
Ne
t p
rofit ($
k)
6% 21% 46% 77% 91%
250 500
% of loans in
size class that
were profitable
The trend in profitability with increasing loan size becomes more apparent
breaking down the portfolio into segments of $100k
21
Profitability
Net profit excluding cost of funds, by loan size (median; interquartile range; increments of USD 100,000)
(1) Net profit = Interest + Fees – credit losses – Operating costs – Recovery costs – Currency losses
Source: CSAF lenders survey conducted between April – June, 2018 of 3,556 individual loan transactions
(*) Sample size < 20
Loan size
$200k
-$50k
$0
$50k
$100k
$150k
$7
00
-80
0k
<$
10
0k
$1
00
-20
0k
$2
00
-30
0k
$6
00
-70
0k
$3
00
-40
0k
$4
00
-50
0k
$5
00
-60
0k
$8
00
-90
0k
$9
00
-$1
M
$1
.0-1
.1M
$1
.1-1
.2M
$1
.2-1
.3M
$2
.0M
+
$1
.3-1
.4M
$1
.8-1
.9M
$1
.4-1
.5M
$1
.9-2
.0M
*
$1
.5-1
.6M
$1
.6-1
.7M
*
$1
.7-1
.8M
Net p
rofit1
($k)
Note that this bar contains
all loans of above $2m,
with the largest loan
represented being for
$10m
In summary, our analyses show that the following ‘risky’ segments were
less profitable, due to three main issues
Small loan sizesAfrica New borrowersLoose value
chains
Longer term
loans
Low income(lower interest
and fee revenue)
High risk(more frequent
and larger credit
losses or
provisions)
High cost(higher operating
costs)
1
2
3
22
Higher
origination
costs
Higher
currency
losses
Lower interest
income
Higher proportion of impaired loans
Higher origination and recovery
costs
The impact of the risky segments compound and drive profitability further
downwards
23
-50
0
50
$30.8k$21.5k
$41.4k
$17.0k$7.9k
$48.6k
$2.3k
$38.5k
$11.9k
-$17.5k
$38.0k
-$39.6k
Average loan transaction revenue, income after operating costs, and income after credit losses
Assuming a 12-month fully-drawn loan of $500k, in USD thousands= 1 risk segment
Loan transaction revenue Income net of operating costs Income net of credit losses
1,562 266 92 188
43% 7% 3% 5%
$706k $767k $649k $326k
# loans1
Rest of World
Existing borrower
Tight value chain
Sub-Saharan Africa
Existing borrower
Tight value chain
Sub-Saharan Africa
New borrower
Tight value chain
Sub-Saharan Africa
New borrower
Loose value chain
% of portfolio1
23% of loans in the CSAF portfolio fall into three or more risky segments; 49% have two or more
(1) Further details in appendix, page 76
Avg. loan size
Agenda
24
1. Introduction
2. Approach
3. Key findings
• Sub-Saharan Africa vs. Rest of the World
• Small vs. large loan sizes
• New vs. existing borrowers
• Loose vs. tight value chains
• Long-term vs. short-term loans
4. Implications
5. Appendix
Loans in Sub-Saharan Africa were less profitable than loans by CSAF
members in the rest of the worldR
es
t o
f W
orl
dS
/Sa
ha
ran
Afr
ica
$14.5k
$29.2k
$29.4k$37.9k
$8.5k
-$20.7k
-$35.1k
$44.2k
$18.3k
Loan
transaction
revenue
$22.3k
Income net
of credit
losses
Credit losses
and recovery
costs
Operating
costs
Income net
of operating
costs
$16.6k
Risk-adjusted
impact cost of
funds
Income net
of cost
of funds
$21.9k
$3.6k
-$13.0k
• Loans in regions outside of Sub-
Saharan Africa were profitable
net of credit losses (before
including costs of funds), while
loans in Sub-Saharan Africa
were not
• The difference was driven by
three key components:
– Lower income from fees and
interest
– Higher operating costs
– A greater amount of credit
losses
25Note: (*) Impact cost of funds used is 3%
Source: CSAF lenders survey conducted between April – June, 2018 of 3,556 individual loan transactions
Currency loss=
Loan economics averages for all CSAF loans analyzed by region
USD thousands
Average
annualized yield
of 0.7% p.a.1
Average
annualized yield
of -4.6% pa.1
(1) For further details, see appendix, page 78
The difference in income was principally driven by greater currency losses
and smaller ticket sizes in Sub-Saharan Africa
7.2
7.3
0 1 2 3 4 5 6 7 8 9
8.4S/Saharan
Africa
7.5Rest of World
Average annualized interest income; Interest income less currency losses (%)
Average annualized fee income; Fee income (%)
1.1
0.7
0.0 0.5 1.0 1.5
S/Saharan
Africa
Rest of World
Currency loss
• While annualized interest income
for Sub-Saharan Africa was
higher than the CSAF average,
currency losses of ~1.2% reduce
the effective interest income to
7.2%, or 0.2 percentage points
lower than the Rest of the World
• However, this difference was
partially compensated for by
greater annualized fee income in
Sub-Saharan Africa
Loan size; Median, interquartile range; $000’s
4002000 100 300 700600500 800 900 1.000
S/Saharan
Africa
Rest of World
• As such, the principal driving
factor for lower income in Sub-
Saharan Africa was the smaller
average loan sizes (median of
$350k vs. $500k for Rest of the
World)
26Source: CSAF lenders survey conducted between April – June, 2018 of 3,556 individual loan transactions
= mean value
Operating costs per loan, particularly origination costs, were higher in
Sub-Saharan Africa than other regions
Operating expenses by region
Average cost per loan, standardized 12 months, USD• Sub-Saharan Africa had the
highest cost per loan
amongst all regions, and
was higher than the Rest of
the World by 22%
• Despite standardizing tenors
to 12 months, overhead and
servicing cost were higher in
Sub-Saharan Africa
• Part of the higher costs may
relate to the higher
proportion of first-time
borrowers and loans in
‘loose’ value chains
$9.9k
$1.9k
$3.2k
$4.1k
$6.7k
Rest of World
$12.9k
$5.0k
$6.5k
Sub-Saharan Africa
$22.6k
$27.6k
+22.1%
Origination costs
Recovery costs
Allocated fixed costs
Servicing costs
27Source: CSAF lenders survey conducted between April – June, 2018 of 3,556 individual loan transactions
Lending in Sub-Saharan Africa had greater annualized credit losses,
principally driven by a higher number of impaired loans for CSAF lenders
• Credit losses were significantly
higher in most loan size brackets
in Sub-Saharan Africa
• Overall, annualized credit losses
in Sub-Saharan Africa were 2.7
percentage points higher than
for the Rest of the World
2.9
4.8
3.43.9
2.1
5.6
8.5
9.7
3.0
4.3
0
1
2
3
4
5
6
7
8
9
10
$500k-1MGrand Total ≤$250k $250-500k >$1M
+90%
Total credit losses
Credit loss (% p.a.), by loan size
Rest of World
S/Saharan Africa
8.4
12.4
8.2
5.26.4
16.6
22.820.8
4.8
10.7
0
5
10
15
20
25
≤$250kGrand Total $250-500k $500k-1M >$1M
+99%
Impaired loans
Percentage of loans, by loan size
• The higher credit losses were
principally driven by a higher
percentage of impaired loans,
rather than by a higher average
value written-off per loan
28Source: CSAF lenders survey conducted between April – June, 2018 of 3,556 individual loan transactions
Agenda
29
1. Introduction
2. Approach
3. Key findings
• Sub-Saharan Africa vs. Rest of the World
• Small vs. large loan sizes
• New vs. existing borrowers
• Loose vs. tight value chains
• Long-term vs. short-term loans
4. Implications
5. Appendix
Large loans were more profitable than small loansL
oa
n s
ize
>$5
00
kL
oa
n s
ize
≤$500k
$20.2k
$6.9k
$24.0k
-$17.9k
$14.1k
-$3.8k
-$24.8k
$51.2k
Credit losses
and recovery
costs
Income net
of operating
costs
$74.8k
Loan
transaction
revenue
Operating
costs
$23.6k
Income net
of credit
losses
$30.1k
$29.2k
Risk-adjusted
impact cost of
funds
Income net
of cost
of funds
$21.0k
-$8.1k
• Loans >$500k were profitable
net of both operating costs and
credit losses, whereas loans of
≤$500k were not profitable net of
operating costs
• The difference was principally
driven by lower income from
small loans, while operating
costs remained the same, and
credit losses were higher as a
percentage of loan size
• Despite their more attractive
economics, larger loans >$500k
were still not profitable after cost
of funds
30Note: (*) Impact cost of funds used is 3%
Source: CSAF lenders survey conducted between April – June, 2018 of 3,556 individual loan transactions
Currency loss=
Loan economics averages for all CSAF loans analyzed by loan size segments
USD thousands
Average
annualized yield
of 2.2% p.a.1
Average
annualized yield
of -8.2% p.a.1
(1) For further details, see appendix, page 79
The difference in income was principally driven by the difference in loan
sizes
8.4
7.0
0 1 2 3 4 5 6 7 8 9
≤$500k 8.6
>$500k 7.4
0.9
0.8
0.0 0.5 1.0
>$500k
≤$500k
Currency loss
• Annualized interest and
annualized fee income were
greater for small loans by 1.4
percentage points and 0.1
percentage points respectively
after currency losses were
accounted for
Loan size; Median, interquartile range; $000’s
0 1,200600 800400200 1,000 1,400 1,600
≤$500k
>$500k
• The principal driving factor for
lower income among small
loans was the smaller ticket
size (median of $267k for
smaller loans vs. $1M for
larger loans)
31Source: CSAF lenders survey conducted between April – June, 2018 of 3,556 individual loan transactions
Average annualized interest income; Interest income less currency losses (% p.a.)
Average annualized fee income; Fee income (% p.a.)
= mean value
Operating costs per loan varied only slightly for loans of different sizes
Operating expenses for an existing by loan size segments
Average cost per loan, standardized 12 months, USD
• Even though revenue
potential varied by loan size,
the operating costs showed
no significant variation (the
slight variations may be a
result of other characteristics
of the loans)
• This results in smaller loans
becoming less profitable than
larger ones$4.4k
$2.3k
$10.3k
$6.3k
$4.4k
≤$250k $250-500k
$2.2k
$10.4k
$6.7k
$2.2k
$11.1k
$4.2k
$23.3k
$6.8k
$500k-1M
$2.0k
$10.6k
$4.1k
$7.0k
>$1M
$23.7k$24.3k $23.7k
+1.9%
Recovery costs
Origination costs
Servicing costs
Allocated fixed costs
32Source: CSAF lenders survey conducted between April – June, 2018 of 3,556 individual loan transactions
Operating costs were largely fixed, while (interest) income depended on
the outstanding balance, making small loans a difficult proposition
33
Months required to cover average operating costs from interest and fee revenue by loan size
Outstanding operating cost, not including recovery costs (months)
If you make a
loan of…
On average only this
amount is actually
earning interest…
And origination fees
will cover this amount
of your operating
costs…
So to break even,
the loan needs to
yield an income
of…
With an average
annualized interest
income after
currency losses of
…
To break even the loan
must be outstanding
for…
$200k $120k 8.6% $22k 8.4% 16 months
$750k $440k 21% $18k 7.6% 3.8 months
$2,000k $1,100k 44% $14k 6.6% 1.2 months
The average tenor of CSAF loans is 14.7 months making it very difficult for small loans to be
profitable even before credit losses or cost of funds are taken into consideration
Credit losses decreased as a percentage of loan size with larger loans
• Credit losses decreased as loan
size increased, with a total
average credit loss of 3.4%
• Up to $1M, decreasing credit
losses were driven by a
decrease in the number of
delinquent loans
• Beyond $1M there was a slight
increase in the proportion of
delinquent loans, but a
simultaneous decrease in the
percentage written-off
3.4
5.9
5.0
3.7
2.4
0
1
2
3
4
5
6
>$1MGrand Total ≤$250k $250-500k $500k-1M
Total credit losses
Credit loss (% p.a.), by loan size
Impaired loans
% of loans impaired; average credit loss for impaired loans, by loan size
8.4
12.4
8.2
5.26.4
0
20
40
60
0
5
10
15
20
≤$250k
% lo
an
s im
pa
ire
d
Avg. c
red
it loss (%
)
Grand Total >$1M$250-500k $500k-1M
Loans in default (%) Average write-off for deliquent loans (%)
34Source: CSAF lenders survey conducted between April – June, 2018 of 3,556 individual loan transactions
Agenda
35
1. Introduction
2. Approach
3. Key findings
• Sub-Saharan Africa vs. Rest of the World
• Small vs. large loan sizes
• New vs. existing borrowers
• Loose vs. tight value chains
• Long-term vs. short-term loans
4. Implications
5. Appendix
Existing borrowers were profitable net of credit losses and operating costs
but not after cost of fundsE
xis
tin
g b
orr
ow
ers
Ne
w b
orr
ow
ers
$36.9k
-$44.9k
$33.4k
$21.0k
$13.0k
$46.5k
-$23.9k
$20.7k
$15.5k
Income net
of cost
of funds
Loan
transaction
revenue
Operating
costs
$41.6k
Income net
of operating
costs
$5.4k
Credit losses
and recovery
costs
Income net
of credit
losses
$14.5k
Risk-adjusted
impact cost of
funds
$20.9k
-$9.2k
• Overall income was similar
between new and existing
borrowers
• Existing borrowers were 1.5x
more profitable as new
borrowers net of operating
costs
• The key drivers of the
difference in profitability were
greater recovery costs and
credit losses as well as higher
operating costs for first-time
borrowers
36Note: (*) Impact cost of funds used is 3%
Source: CSAF lenders survey conducted between April – June, 2018 of 3,556 individual loan transactions
Loan economics averages for all CSAF loans analyzed by borrower type
USD thousands
Average
annualized yield
of 1.1% p.a.1
Average
annualized yield
of -3.6% p.a.1
(1) For further details, see appendix, page 80
Lenders had higher origination costs for a new borrower compared to
existing borrowers
Operating expenses for an existing borrower vs. a new borrower
Average cost per loan, standardized 12 months, USD• A first-time borrower had
nearly 50% higher costs than
an existing borrower
• 7 out of 9 lenders reported
higher origination costs for
first-time borrowers of 1.5x
on average
• The higher risk associated
with a first-time borrower
also had significantly higher
recovery costs than an
existing borrower
$5.3k
$1.6k $14.8k
$4.0k
$9.2k
$6.5k
Existing borrower
$4.2k
$7.3k
New borrower
$21.2k
$31.6k
+48.7%
Recovery costs
Origination costs
Servicing costs
Allocated fixed costs
37Source: CSAF lenders survey conducted between April – June, 2018 of 3,556 individual loan transactions
First-time borrowers had greater annualized credit losses, principally
driven by a higher percentage of impaired loans
• Credit losses were significantly
higher in most loan size brackets
among first time borrowers
• Overall, annualized credit losses
among new borrowers were
74% higher than among existing
borrowers
2.8
5.2
4.4
3.1
1.8
4.9
7.4
6.15.5
3.9
0
1
2
3
4
5
6
7
8
$500k-1MGrand Total $250-500k≤$250k >$1M
+74%
Total credit losses
Credit loss (% p.a.), by loan size
Existing borrower
First time borrower
6.7
11.1
6.7
3.95.0
13.414.8
12.511.6
14.0
0
5
10
15
Grand Total ≤$250k $500k-1M$250-500k >$1M
+101%
Impaired loans
Percentage of loans, by loan size
• The higher credit losses were
principally driven by a higher
percentage of impaired loans
38Source: CSAF lenders survey conducted between April – June, 2018 of 3,556 individual loan transactions
Agenda
39
1. Introduction
2. Approach
3. Key findings
• Sub-Saharan Africa vs. Rest of the World
• Small vs. large loan sizes
• New vs. existing borrowers
• Loose vs. tight value chains
• Long-term vs. short-term loans
4. Implications
5. Appendix
Tight value chains were more profitable despite having lower fee and
interest revenue than loose value chainsC
off
ee
/Co
co
a
Loan economics averages for all CSAF loans analyzed by value chain group
USD thousands
Oth
er1
$25.5k
$31.2k
$32.9k
-$31.8k
$57.9k
$26.7k
-$6.2k
Loan
transaction
revenue
Credit losses
and recovery
costs
Risk-adjusted
impact cost of
funds
$19.8k
$14.1k
Operating
costs
Income net
of operating
costs
Income net
of credit
losses
$11.0k
Income net
of cost
of funds
$0.6k
$34.6k
$14.7k
-$10.4k
• Loans to coffee and cocoa were
profitable net of credit losses,
while loans to loose value chains
were not
• Coffee/Cocoa had lower income,
but overall profitability was
higher due to lower operating
costs, and lower recoveries and
credit losses
• Both segments made losses
after cost of funds was taken
into account
40Note: (*) Impact cost of funds used is 3%
Source: CSAF lenders survey conducted between April – June, 2018 of 3,556 individual loan transactions
(1) Further details of profitability by additional value chains in appendix, page 66
(2) For further details, see appendix, page 81
Average
annualized
yield of 0.2%2
Average
annualized
yield of -0.8%2
The difference in income was principally driven by the shorter tenor of
loans in the coffee and cocoa value chain
8.0
6.7
0 1 2 3 4 5 6 7 8 9
Coffee/Cocoa
Other
8.4
7.1
Average annualized interest income; Interest income less currency losses (% p.a.)
Average annualized fee income; Fee income (% p.a.)
1.2
0.5
0.0 0.5 1.0 1.5
Other
Coffee/Cocoa
Currency loss
• Annualized interest and
annualized fee income was
greater for coffee/cocoa by 1.4
percentage points and 0.7
percentage points after
currency losses were taken
into account
Tenor length; Median, interquartile range (months)
0 2 4 6 8 10 12 14 16 18 20 22
Other
Coffee/Cocoa
• While loans for coffee/cocoa
were slightly higher on average
(median of $500k vs $420k for
loose value chains) income
was lower due to the shorter
loan tenor
41Source: CSAF lenders survey conducted between April – June, 2018 of 3,556 individual loan transactions
Lenders had higher origination and recovery costs for loans outside the
coffee and cocoa value chains
Operating expenses for a coffee and cocoa vs. other value chains
Average cost per loan, standardized 12 months, USD• Loans in the other value chains
had over 30% higher costs than
coffee and cocoa
• 5 out of 9 lenders reported higher
origination costs for non-coffee
and cocoa loans of an average of
1.5x
• The additional risk associated with
loose value chains also has
significantly higher recovery costs
than an existing borrow
• This higher cost and risk may
disincentivize lenders to lend in
value chains other than coffee and
cocoa
$4.4k
$1.6k
Other
$3.4k
$8.7k
$6.7k
Cocoa & Coffee
$14.0k
$4.2k
$6.7k
$21.3k
$28.3k
+32.5%
Recovery costs
Origination costs
Servicing costs
Allocated fixed costs
42Source: CSAF lenders survey conducted between April – June, 2018 of 3,556 individual loan transactions
Higher credit losses outside of the coffee/cocoa value chain were driven
by a higher proportion of impaired loans
• Credit losses were higher outside
of the coffee and cocoa value
chains among all loan brackets
apart from loans of greater than
$1M
• Overall, annualized credit losses
were 0.2 percentage points higher
than in the coffee/cocoa value
chain
3.3
4.94.3
2.9 2.8
3.5
6.8
5.7
4.5
2.2
0
1
2
3
4
5
6
7
≤$250kGrand Total $250-500k >$1M$500k-1M
+7%
Total credit losses
Credit loss (% p.a.), by loan size
Coffee/Cocoa
Other
• The higher credit losses were
driven by a higher percentage of
impaired loans for loan sizes up to
$1M
Impaired loans
Percentage of loans (%), by loan size
43Source: CSAF lenders survey conducted between April – June, 2018 of 3,556 individual loan transactions
5.57.1 6.1
3.6 4.4
13.6
20.7
12.3
8.0
11.1
0
5
10
15
20
25
>$1MGrand Total ≤$250k $250-500k $500k-1M
+149%
Agenda
44
1. Introduction
2. Approach
3. Key findings
• Sub-Saharan Africa vs. Rest of the World
• Small vs. large loan sizes
• New vs. existing borrowers
• Loose vs. tight value chains
• Long-term vs. short-term loans
4. Implications
5. Appendix
Short-term loans were more profitable than long-term loans net of credit
losses, though not after cost of fundsTe
no
r ≤1
2 m
on
ths
Te
no
r >
12
mo
nth
s
$48.6k
$42.1k
$81.1k
$37.8k
$38.9k
-$9.7k
-$47.4k
$8.5k
Risk-adjusted
impact cost of
funds
$17.4k
Loan
transaction
revenue
Operating
costs
Income net
of operating
costs
Income net
of credit
losses
$9.9k
-$6.4k
Credit losses
and recovery
costs
Income net
of cost
of funds
$29.4k
$12.0k
$2.1k
• Loans with up to12-month tenors
were profitable net of credit
losses whereas longer terms
loans were not
• The difference in profitability was
driven by lower income among
short-term loans, as well as
higher operating, credit loss and
recovery costs
45Note: (*) Impact cost of funds used is 3%
Source: CSAF lenders survey conducted between April – June, 2018 of 3,556 individual loan transactions
Loan economics averages for all CSAF loans analyzed by tenor segments
USD thousands
Average
annualized yield
of 0.6% p.a.1
Average
annualized yield
of -1.1% p.a.1
(1) For further details, see appendix, page 82
Total interest and fee income was lower among short-term loans despite
higher annualized income
8.3
6.5
0 1 2 3 4 5 6 7 8 9
7.2
8.3≤12 months
>12 months
Average annualized interest income; Interest income less currency losses (% p.a.)
Average annualized fee income; Fee income (% p.a.)
1.2
0.5
0.0 0.5 1.0 1.5
≤12 months
>12 months
Currency loss
• Annualized interest and
annualized fee income was
greater for short-term loans by
1.8 percentage points and 0.7
percentage points lower
respectively after currency
losses were taken into account
Tenor length; Median, interquartile range (months)
0 5 10 15 20 25 30 35 40 45 50 55
≤12 months
>12 months
• The principal driving factor for
lower income was the shorter
tenor
46Source: CSAF lenders survey conducted between April – June, 2018 of 3,556 individual loan transactions
Operating costs per loan were higher for longer-term loans after adjusting
for tenor
Operating expenses by tenor
Average cost per loan, standardized 12 months, USD
• The longer-term loans
showed reasonably higher
operating costs even when
adjusted to average over a
12-month period, owing
largely to the higher recovery
costs
• There was also a higher
origination cost associated
with the longer tenor loans,
potentially owing to higher
diligence efforts required for
longer-term lending
$4.8k
$6.5k
$1.4k
≤12 months
$4.3k
$10.2k
$28.0k
$11.7k
$7.2k
$4.3k
>12 months
$22.3k
+25.7%
Recovery costs
Origination costs
Allocated fixed costs
Servicing costs
47Source: CSAF lenders survey conducted between April – June, 2018 of 3,556 individual loan transactions
Higher credit losses among long-term loans were principally driven by a
higher percentage of impaired loans
• Credit losses were significantly
higher in most size brackets
among long-tenor loans
• Overall, annualized credit losses
for longer-term loans were 1.1
percentage points higher than
among short-term loans
2.8
4.1 4.0
2.12.7
3.9
7.0
5.8 5.8
2.3
0
1
2
3
4
5
6
7
Grand Total >$1M$250-500k≤$250k $500k-1M
+38%
Total credit losses
Credit loss (% p.a.), by loan size Tenor ≤12 months
Tenor >12 months
Impaired loans
Percentage of loans (%), by loan size
48Source: CSAF lenders survey conducted between April – June, 2018 of 3,556 individual loan transactions
4.16.2
3.92.4
4.0
20.3
24.8
21.8
16.3
12.9
0
5
10
15
20
25
>$1MGrand Total ≤$250k $500k-1M$250-500k
+393%
• The higher credit losses were
principally driven by a higher
percentage of impaired loans
Agenda
49
1. Introduction
2. Approach
3. Key findings
4. Implications
5. Appendix
There are several ways in which donors can address the finance gap for ag-
SMEs using blended finance instruments and other supporting mechanisms
Blended
finance
instruments
Other
supporting
mechanisms
Output-based
incentives
Risk mitigation
Direct funding
Technical
assistance
Cost-cutting
technology and
innovationCoordinated
value chain
interventions
Enabling
environment
Instruments that incentivize private investment in high-impact but
underdeveloped segments of the market
Mechanisms that protect investors against systemic risks to their
portfolio, or reduce potential losses if risks materialize
Interventions to increase addressable demand by building capacity for
prospective borrowers and/or also providing supporting to lenders to
hone their financial products and delivery to reach the market
Direct concessional funding in a financial service provider serving
high-impact and underdeveloped market segments
Support disruptive technological innovations; encouraging new actors
to enter the market with different business models to serve agricultural
SMEs, thereby driving competition and efficiency
Support interventions that increase coordination throughout the value
chain to enhance efficiency and transparency
Engage public and private sector actors to identify and address legal,
regulatory, and policy barriers; target key infrastructure investments;
and facilitate dialogue and learning exchange
1 2 3
1 2 3
1 2 3
1 2 3
Driver
addressed
1 2 3
1 2 3
1 2 3
50
High risk1 2 3Low income High cost
Output-based incentives: A pay-per-loan facility could allow lenders
to increase profitability on low-margin borrowers
51
Objective• Increase addressable lending market by incentivizing loans in high-impact but underserved segments
• Make low-margin / loss-making loans more economically viable for lenders
How it would
work
Lenders receive additional revenue for each loan made within specified segments or meeting certain impact
criteria
• Committee sets eligibility criteria for impact segments, e.g.,:
– Frontier markets: loans in sub-Saharan Africa
– Small facility size: loans under $500k
– High-cost value chains: loans in loose value chains
• Lenders that make qualifying loans can submit funding requests to the committee on a periodic basis (e.g.,
annual or semi-annual)
• Committee reviews applications and provides funding for all qualifying loans
Risks &
mitigation
• Potential risk of lenders misrepresenting loan classifications, or gamifying facility sizes to maximize reward
Mitigation: clearly defined criteria parameters and caps; milestone-based reward design; random audits
• High degree of overhead may be required to assess payouts to lenders at a loan-level
Mitigation: outsourcing to 3rd parties specialized in verification
Theory of
Change
Supporting earlier-stage, smaller borrowers on their growth journey with a time-bound subsidy can help those
such borrowers grow their enterprises in a way that makes them profitable to serve by CSAF lenders or other
intermediaries
For example a tiered-system could use a points-based approach to
determine the required level of subsidy
52
Size Number
of loans
Borrowers
status
Value
chain
type
Average loan
size
Average
subsidy need
per loan (%)1
Average
subsidy need
per loan
(USD)
Implied donor
leverage2
1st tier >500k 1223 existing - $1.2M N/A N/A N/A
2nd tier >500k 248 new - $1.3M 0.2% $5k 260x
3rd tier
≤500k
1781
existing -
$0.3M 6.4% $20k 15x
≤500k new tight
4th tier ≤500k 304 new loose $0.2M 16.8% $40k 5x
(1) Based on CSAF lenders data only
(2) Calculation: Average loan size / Average subsidy per loan
Source: CSAF lenders survey conducted between April – June, 2018 of 3,556 individual loan transactions
ILLUSTRATIVE EXAMPLE
Risk mitigation: A first-loss protection facility could make
lending in riskier segments more attractive for lenders
• Allow testing of lending segments with high perceived risks and identify potential pockets of profitability
• Increase inclusion by providing access to borrowers without prior credit history
Lenders lending in ‘risky’ loan segments are eligible for access to a credit guarantee facility to provide first loss
buffers and/or cover losses (partial or full) in case of default by borrower
• Committee sets eligibility criteria for high-risk and high-impact segments, as well as parameters levels of
loss cover for each segment
– First-time borrowers: lending to an SME without prior credit history
– Long-term / asset-finance borrowing: lending to an SME with a 12 month+ tenor for repayment
• Lenders can submit details of their portfolio within the sub-segment, of which the first-loss facility can be
utilized to the extent of the % cover offered
• Committee reviews applications and provides credit loss cover based on defined parameters
• Potential to create moral hazard of encouraging perverse risk-taking behavior of lenders
Mitigation: first-loss facility at a portfolio level, if provided by generating a high volume of qualifying loans;
careful definition of parameters, as well as ensuring appropriate caps on the facility;
53
Objective
How it would
work
Risks &
mitigation
Theory of
Change
Support market development of segments perceived to be risky to test and identify pockets for sustainable
lending and attract commercial capital to high-impact segments through a demonstration effect
An example estimate of how much a first-loss cover facility would
require to subsidize based on the CSAF portfolio
Risk segment
% impaired
within risk
segments
% impairment
for non-risk
segment1
First-loss cover
for incremental
risk %
Total guarantee cover
for $10M incremental
lending
Frontier markets (Sub-
Saharan Africa)7.1% 2.7% 4.4% $440k
First-time borrowers 6.8% 2.7% 4.1% $410k
Long-term loans (>12
months)9.2% 1.7% 7.5% $750k
Small facility loans (<500k) 5.0% 2.4% 2.6% $260k
Loose value chain loans
(non-coffee / cocoa)6.3% 2.3% 4.0% $400k
54
(1) Respectively, rest of the world, existing borrowers, short term loans, big facility size and tight value
chain loans
ILLUSTRATIVE EXAMPLE
Source: CSAF lenders survey conducted between April – June, 2018 of 3,556 individual loan transactions
Direct funding: Balance sheet support would allow lenders to
take on more risk while meeting investor requirements
• Allow lending in riskier segments while meeting capital requirements around risk covenants
• Increase access to long-term loans for SMEs to fuel their longer-term investments and growth
Lenders would receive balance sheet support to improve leverage ratio and increase risk appetite for making
higher risk loans without breaching investor covenants or requirements
• Size of balance sheet support directly proportional to increase in ‘higher risk’ capital available
• Alternatively, concessional interest rates could be offered to lower average cost of funds
• Lenders that demonstrate higher impact and willingness to take on more difficult segments would receive
continued support, others would see balance sheet support decrease over time
– Frontier markets: loans to SMEs in sub-Saharan Africa
– Loose value chains: loans in value chains other than coffee or cocoa
– Long-term / asset-finance borrowing: lending to an SME with a 12 month+ tenor for repayment
• Potential to create moral hazard of encouraging perverse risk-taking behavior of lenders
Mitigation: careful definition of parameters, as well as ensuring appropriate caps on the facility
55
Objective
How it would
work
Risks &
mitigation
Theory of
Change
Providing balance sheet support will allow lenders to have higher exposure to risk while meeting risk
covenants, and a lower cost of capital its cost of capital
An example of how direct funding could increase lender’s risk
appetite, by providing 10% buffer capital
56
20%10%
80%
15%
75%
Including grant-
supported funding
Standard funding stack
Class A - Senior debt
Class C - Grant
Class B - Catalytic debt
5.44.8
7.6
16.8
Standard funding Grant supported funding
Cost of funds
Risk covenant
Illustrative standard and grant supported funding
structure (%)Illustrative Total cost of funds and risk covenant
for standard and grant- supported funding
structures (%)
(*) Risk covenant defined as the maximum percentage of lent capital that can be at risk
Cost of
funds
Risk
covenant*
5% 7%
7% 10%
0% 100%
ILLUSTRATIVE EXAMPLE
Intermediaries
with financial
buffer would
have lower return
and higher risk
thresholds
compared to
standard funding
structures
Technical assistance: Advisory services can be provided to lenders
and borrowers to lower costs and risk for agricultural SME finance
• Provide services to lenders to lower transaction costs and/or improve risk management
• Provide services to borrowers to improve enterprise management capacity and lower lending risk
Lenders would have access to pooled resources such as standardized templates, TA facilities, deal
flow/match-making platform, etc. and/or borrowers would have access to advisory services which could
increase their performance and reduce risk
• Use would be only for member organizations that qualify to meet certain eligibility requirements, e.g.,
– Loan volume: Must have certain number of transactions to qualify for access to TA facility
• Standardized pricing and other guidance on how and when to use would be provided
• Potential of providing high-cost initiatives which may not yield desired results or outcomes
Mitigation: conducting demand assessment; running pilot programs
57
Objective
How it would
work
Risks &
mitigation
Theory of
Change
Providing assistance to lenders can reduce some transaction costs to make lending in low-revenue segments
profitable, while support to borrowers reduces risk of lending to them
Some examples of assistance programs for lender and borrower
technical assistance can be explored
• Technical assistance can be provided to
borrowers to increase lender confidence, such
as:
– Improving financial reporting and accounting
procedures
– Strengthening governance mechanisms
• Training support for borrowers may include
advisory services on business planning and
growth, such as:
– Growth / expansion strategies
– Process efficiency
58
Borrower assistanceLender assistance
• Support transaction process through advisory
services, including:
– Pipeline building and origination
– Due diligence support
– Legal service providers
• Training facilities may include support on
gaining internal efficiencies, such as:
– Streamlining of processes
– Capacity building of human capital
– IT system development
Cost-cutting technology and innovation: encouraging role of new
disruptive actors in the ecosystem to increase efficiencies
Description
Supporting the creation and catalytic technology and business model innovations to bring in
new actors and drive efficiencies in the market in the form of reduced cost to serve and
increased transparency
• Financial services:
– Alternative credit scoring and credit monitoring to bring new customer segments into the addressable
market for input and post-harvest financing
– Agri insurance that leverages more localized data on farms and the surrounding conditions (e.g.,
through satellite technology or drones), helping insurers more easily manage risk and analyze premiums
– Cashless payments that eliminates the risk of holding and transporting cash
• Shared services providers:
– Centralized legal services that ensure quality and lower costs for individual lenders
– 3rd party monitoring officers that can conduct routine activities at a local level
Examples
59
Coordinated value chain interventions: create a more attractive
lending market by increasing market efficiency and transparency
• Supply chain management:
– Product tracing: Digital platform for tracking and quality assurance, from certification to quality declared (e.g.,
barcode or scratch coin verification for seeds)
– Logistics coordination: Digital platforms to link farmers/traders to available storage facilities
• Market access:
– Aggregation tools: Digital platform for tracking and quality assurance, from certification to quality declared
(e.g., barcode or scratch coin verification for seeds)
– Digital marketplaces: virtual cooperatives that match it to supply, provide real time price information on inputs,
and facilitate low-cost transactions between farmers and input providers
• Ag intelligence, knowledge, and management:
– Ag trends mapping and prediction: Harmonized, digital database to manage regulatory systems
60
Supporting programs and initiatives along agriculture value chain that allow increased
coordination, efficiency and transparency, to encourage additional lending activity from new
and existing playersDescription
Examples
Enabling environment: encourage policy and infrastructure
decisions that support agri-SMEs
• Identify legal, regulatory, and policy barriers: conduct legal and regulatory assessment to
identify barriers to agri-SME financing and broader sector development
• Engage partners and influencers: collaborate with other actors (e.g., NGOs, multi-lateral / bi-
lateral organizations, policy organizations) in the market to align agendas and pave a
common path forward to influence policymakers
• Develop communities to share knowledge: engage different intermediaries the sector in
various forums that encourage to knowledge sharing and idea generation
61
Supporting local initiatives of public and private sector actors on enhancement of policy,
infrastructure and knowledge sharing Description
Examples
In summary, these tools can both, support existing lenders to expand their
addressable market, and attract new actors in the sector
Blended
finance
instruments
Other
supporting
mechanisms
Output-based
incentives
Risk mitigation
Direct funding
Technical
assistance
Cost-cutting
technology and
innovationCoordinated
value chain
interventions
Enabling
environment
• Support existing lenders to continue
lending sustainably while expanding
addressable market to high-impact, hard-
to-serve agricultural SME segments
• Encourage new players and innovative
business models to serve agricultural
SMEs
• Create an ecosystem that can attract
more lenders and more capital in the
underserved agricultural SME lending
space
1 2 3
1 2 3
1 2 3
1 2 3
Driver
addressed
1 2 3
1 2 3
1 2 3
Expected outcomes and impact1 2 3Low income High cost High risk
Agenda
63
1. Introduction
2. Approach
3. Key findings
4. Implications
5. Appendix
Appendix 1: Research confirmed the significant social impact of bridging
the agricultural SME financing gap
• The agricultural sector has huge lever on poverty: three-quarters of the developing world lives in rural areas, and about nine out of
every ten depend upon agriculture for their livelihoods1. Agricultural investment is often regarded as one of the most efficient and effective
ways to promote food security and reduce poverty, with some studies demonstrating a four-fold reduction in poverty over other sectors2.
• Evidence suggests improved access to finance products for agricultural SMEs can help boost smallholder farmers consumption, food
security, income, production, and resilience to external schocks3. It also suggests a wide variation of take-up for financial products and
positive effects concentrated on certain pockets of borrowers. Results from studies that examined agriculture-specific products suggested
positive impacts on production, use of formal financial services, and maize inventories.
• There are two principal channels through which credit interventions impact rural populations:
– Access to credit could allow farmers to invest in agricultural inputs such as labor, land area cultivation, equipment, improved variety seeds, or
fertilizers which they might not otherwise be able to afford. Increased investment in inputs should lead to increased production.
– Access to credit could give poor households another strategy to cope with risk: in the case of a shock, the household could borrow money rather
than liquefying assets, limiting consumption, or selling off-farm labor.
• Two examples:
– A World Bank policy research working paper4 measured the impacts of semi-formal credit provided by cooperatives, input suppliers, microfinance
institutions, and NGOs in rural Rwanda. Removing all household-level credit constraints was estimated to increase the total value of a household’s
agricultural output by 17 percentage points, from USD 272 to USD 326.
– A 2014 randomized control trial study in a rural area of Morocco dominated by smallholder agriculture found an average 140 percent return to
microcredit capital on business profit.
64
(1) Oxfam. 2009. The Missing Middle in Agricultural Finance — Relieving the Capital Constraint on Smallholder
Groups and Other Agricultural SMEs. (2) World bank. 2007. World Development Report 2008: Agriculture for
Development. (3) Rural Agricultural Finance Learning Lab. 2015. Evidence on the Impact of Rural and Agricultural
Finance on Clients in Sub-Saharan Africa: a Literature Review. (4) World bank. Policy research working paper. 2014.
Credit constraints, agricultural productivity, and rural nonfarm participation : evidence from Rwanda (4) Crepon, et al.
2014. NBER working paper. Estimating the Impact of Microcredit on Those Who Take It Up: Evidence from a
Randomized Experiment in Morocco
Appendix 2: Detailed value chain breakdown of loan performance (1/2)
65
Variable Coffee CocoaCashew
nutsQuinoa Honey Sesame Brazil nuts Soy Other nuts
Macadamia
nutsOther crops
# loans 2023 280 114 78 67 65 43 43 38 37 768
Av. loan size ($k) 663 046 758 381 718 242 956 710 447 329 723 487 713 951 583 641 565 601 707 090 616 855
Av. Tenor (months) 11.9 12.1 11.9 7.7 11.9 12.5 10.7 11.0 10.1 12.6 24.7
Interest yield p.a.* (%) 8.5% 7.8% 8.2% 7.9% 9.0% 8.3% 8.1% 8.9% 7.4% 8.9% 6.7%
Fee yield p.a.* (%) 1.2% 1.0% 0.7% 0.4% 0.9% 0.7% 0.6% 0.5% 0.6% 1.6% 0.4%
Write-off p.a.* (%) 3.2% 3.9% 2.5% 4.0% 0.7% 9.1% 0.0% 2.4% 8.7% 0.4% 3.5%
Currency losses p.a.*
(%)0.01% 1.65% 0.00% 0.00% 0.00% 0.04% 0.00% 0.00% 0.02% 0.00% 0.57%
Op. cost/loan ($k) 19 909 19 359 22 583 19 919 23 406 18 997 22 313 21 624 15 386 27 457 37 325
Recovery cost/loan
($k)1 591 2 675 2 373 1 996 918 6 755 1 942 4 636 3 301 2 724 4 798
Cost of funds/loan ($k) 10 353 15 709 16 855 11 458 8 820 17 683 13 910 11 689 14 879 14 375 32 859
Net profit = Interest + Fees – credit losses – Operating costs – Recovery costs – Currency losses. (*) Annualized values
Appendix 2: Detailed value chain breakdown of loan performance (2/2)
66
Macadamia
nuts
-26.3
QuinoaHoneyBrazil
nuts
Cashew
nuts
18.2
Soy Cocoa
-10.8
Coffee Nuts Sesame Other
crops
17.6
12.9
3.9 3.81.3
-3.4 -4.0
-22.3
Profitability
Net profit, by value chain (USD thousands)
Net profit = Interest + Fees – credit losses – Operating costs – Recovery costs – Currency losses. (*) Annualized values
# Loans: 43 37 114 67 43 2023 78 280 38 65 768
Average: -2.4
Macadamia, Brazil
and Cashew nuts
were the most
profitable crops while
coffee was only
marginally profitable
Appendix 3: Loan data collection methodology (1/7)
• Dalberg requested data on value chain, country, first time borrower status, product type,
loan/facility size, origination fees, interest type, annual interest rate (%), date of first
disbursement, currency and balance at 30+, 90+, 180+ and 365+ days past due (dpd)
• Dalberg analyzed data on 3,556 loans out of 4,488 received from nine CSAF lenders
– Dalberg omitted 391 loans (11% of loans) as they were out-of-scope (non-agricultural
value chain, outside of time period under consideration, or in a region other than Africa,
Asia or Latin America)
– Dalberg excluded a further 284 loans (7.1% of remaining loans) without loan
characteristics data
– Dalberg excluded a further 252 loans (6.6% of remaining loans) with incomplete or
inconsistent loan transaction data
Data
collection
summary
Data
limitations
• Some lenders provided loan-level transaction data segmented by customer. In some cases this
resulted in overlapping loans to one customer. To standardized these, Dalberg separated loans
to the same customer into separate segments (or borrower transaction relationships) where
there existed an interim period in which the outstanding balance fell to zero
• Some lenders provided slightly incomplete loan characteristics data (e.g. missing contractual
interest rate, tenor, etc.) for a subset of loans with corresponding loan transaction data. Where
this was the case, Dalberg omitted these loans for outputs in which they were directly relevant,
but included them in all other analyses
67
Collection
Appendix 3: Operating costs data collection methodology (2/7)
68
Data
collection
summary
Data
limitations
• Dalberg requested each lender to share their operating cost data, including personnel costs,
travel, legal / professional, and other
• Where possible, Dalberg also requested cost breakdown for Headquarters vs. business units in
different regions
• Additionally, Dalberg asked each lender to estimate the percentage of their total costs they
would allocate to specifically their agri-lending activities
• Finally, during individual lender meetings and through follow-ups after, Dalberg requested
allocation of costs across the different stages of the loan lifecycle
• 4 out of the 9 lenders shared data in the standardized format requested in the template, while
the others shared cost data in other formats used internally or in financial reports
• Different organizations have varying methods of reporting operating costs that are difficult to
disaggregate and allocate with full accuracy
• Variations in organization structures and locations can have significant impact on their operating
costs
Collection
69
• Dalberg standardized the start date of all loans to calculate the utilization and duration of the loans
• Dalberg standardized the product type to either working capital or asset finance. Value chain was assigned in a
hierarchical structure: (i) we assigned loans in the coffee or cocoa value chain as coffee or cocoa; (ii) for remaining
loans, we assigned loans in hard currency to ‘export oriented’, and loans in local currency to ‘domestic oriented’.
We classified size segment based on the loan/facility size variable
• The value of disbursements, repayments, write-offs, fees, interest and outstanding balance was converted to USD
(or EUR for West African Francs) using exchange rate data with month-level granularity. For loans analyzed in
EUR, this report presents total results in USD converted at today’s exchange rate to enable a like-for-like
comparison
• A weighting factor – the dollar duration - was calculated as the product of loan size with loan duration in years. We
then calculated annualized interest yield p.a., annualized fee yield p.a., annualized write-off yield p.a. and
annualized currency loss percentage as the value of the variable in dollars divided by the weighting factor (dollar-
duration) of the loan1
• For currency losses for loans in local currency, the total currency loss was calculated as the total disbursed (USD)
– total repaid (USD) – total write-off (USD) – total outstanding balance (current USD)
• For loans in delinquency, in the absence of lender guidance on expected recovery, Dalberg assumed 0% recovery
for active loans 365+ days past due (DPD); 25% for 180–365 DPD; 50% for 90–180 DPD; and 75% for 30–90 DPD
• Where write-offs were listed beyond two years after the last disbursement or repayment, Dalberg applied a cut
such that the write-off was registered 24 months after the last disbursement or repayment. This was done to
prevent an unfairly high weighting on loans with a long time to write-off in aggregated figures
Standardization
1) For aggregated figures, the summed values of the loan variable was divided by the summed values of
the weighting factor
Appendix 3: Loan data standardization and analysis (3/7)
70
Operating
cost data
standard-
ization
• Dalberg assigned CSAF lender operating costs for lending activity to the three stages of the
lending process, and validated them through one-on-one meetings with each lender
– Annual operating costs for each lender were allocated to the different stages of the loan
lifecycle based on estimations from lenders around time and effort across each
– Costs associated with non-agriculture lending, technical assistance, and fundraising
activities were excluded from operating costs
– Costs were adjusted for regional-level differences within each lender
– Costs were then allocated at a loan level by stage:
o Origination costs were allocated based on the year of origination
o Servicing and overhead costs were allocated based on month and years of active tenor
o Recovery costs were allocated over the period analyzed
– Manipulations for unavailable data:
o Extrapolated data breakdown for years based on financial statements
o Allocated regional costs based on lending activity
o Extrapolated 2017 per loan costs for servicing and overheads loans with tenors beyond 2017
– Exclusions:
o Dalberg excluded loans of lenders' first year of operations as costs allocated to an individual loan
would be overstated
Standardization
Appendix 3: Operating costs data standardization (4/7)
71
1) Each loan is assigned the appropriate cost from each year that it is active, based on how long (i.e.,
number of months) the loan is active that year
2) Recovery costs assigned as lifetime cost of a loan to its vintage (i.e., year of origination)
Active period is determined based on date of disbursement and contractual tenors
• Sourcing / Pipeline development
• Due diligence
• Processing and underwriting
𝑂𝑟𝑖𝑔𝑖𝑛𝑎𝑡𝑖𝑜𝑛 𝑐𝑜𝑠𝑡s 𝑓𝑜𝑟 𝑡ℎ𝑒 𝑦𝑒𝑎𝑟
#𝑙𝑜𝑎𝑛𝑠 𝒐𝒓𝒊𝒈𝒊𝒏𝒂𝒕𝒆𝒅 𝑤𝑖𝑡ℎ𝑖𝑛 𝑡ℎ𝑒 𝑦𝑒𝑎𝑟
2. Servicing
• Monitoring
• Reporting
• Client Servicing
IT / Admin / Backoffice
HQ allocation
𝑆𝑒𝑟𝑣𝑖𝑐𝑖𝑛𝑔 𝑐𝑜𝑠𝑡 𝑓𝑜𝑟 𝑡ℎ𝑒 𝑦𝑒𝑎𝑟
𝐹𝑢𝑙𝑙 𝑦𝑒𝑎𝑟 𝒆𝒒𝒖𝒊𝒗𝒂𝒍𝒆𝒏𝒕 𝒍𝒐𝒂𝒏𝒔 𝒂𝒄𝒕𝒊𝒗𝒆𝑖𝑛 𝑡ℎ𝑒 𝑦𝑒𝑎𝑟1
3. Recovery
• Collection
• Legal recourse
• Collateral liquidation
Salaries
Travel
Legal Legal
𝑇𝑜𝑡𝑎𝑙 𝑟𝑒𝑐𝑜𝑣𝑒𝑟𝑦 𝑐𝑜𝑠𝑡𝑠 𝑖𝑛 𝑜𝑟𝑔 ℎ𝑖𝑠𝑡𝑜𝑟𝑦
# 𝑙𝑜𝑎𝑛𝑠 𝒊𝒏 𝒓𝒆𝒄𝒐𝒗𝒆𝒓𝒚 𝑎𝑡 𝑎𝑛𝑦 𝑝𝑜𝑖𝑛𝑡2
Activities
Direct costs
Indirect costs
(overheads)
Loan-level
allocation
1. Origination
Standardization
Appendix 3: Operating costs data analysis (5/7)
72
• Dalberg designed its cost of funds approach to reflect the specific condition of institutions facing impact driven
investors and lenders. The two main elements of this analysis were:
– Average COF was set at 3% after discussion with CSAF lenders
– This average was risk adjusted on a loan-by-loan basis, based on loan characteristics data and sample
averages
• Dalberg conducted a risk adjusted estimation of cost of funds on a loan by loan by loan basis. Dalberg made the
following assumptions:
– Debt to equity ratio. Based on Basel III Advanced IRB risk-weighted assets formula. The probability of
default was estimated based on regional and size segment averages observed in our data, using linear
extrapolation for Asia and East Africa outliers. Loss given default was set at 75%
– Cost of debt. Calculated by loan as the sum of risk free rate and a risk premium. Dalberg used EUR yield
curves for EUR and XOF and USD yield curves for all other currencies
– Cost of equity. Calculated for all loans based on a 20% average ROE. Average calculated for a
representative panel of largest banks in each major CSAF country.
• Dalberg normalized the results to average at 3% p.a. with 0.5% standard deviation.
Analysis
Appendix 3: Cost of funds analysis (6/7)
Appendix 3 : Annualized yield calculation (7/7)
73
Dollar duration ($)
Annualized yield p.a. (%
per $ per year)
= duration x loan size
12
= Total porfolio income
Total porfolio dollar duration
An example portfolio, composed of 4 loans
Average balance over time, USD
Total area = area (1) + area (2) + area (3) = 3
Total portfolio credit loss = 300k
1.6M
2 years
4M
500k
1 year
Loss = 300k
1
2
3
3M
1 year
One year equivalent to the example portfolio
Average balance over time, USD
Total area = 3 = 1 year x dollar duration
Total portfolio dollar duration = $3M
A portfolio with a dollar duration of $100 is equivalent to having one
loan of $100 fully outstanding for 1 year
Illustration : annualized
credit-loss %=
Total porfolio credit loss
Total porfolio dollar duration=
$300k
$3M= 10%
A portfolio with an annualized income yield p.a. of 1% and an
average balance of $100 over a year will bring $1 of income over
the year
A portfolio with an annualized credit loss percentage of 10% and an
average balance of $3M over a year will incur a loss of $600k over the
year
6 months
Appendix 4 : All CSAF lenders data by geography and value chain (1/2)
1. Based on weighted average of individual loan yields by dollar duration.
2. Loans where no interest rate was provided have been omitted for this variable
Region Value chain # loansAverage loan
sizeHeadline
Interest Rate2
Interest yield p.a. (annualized,
gross) 1
Fee yield p.a. (annualized) 1
credit losses, net of recoveries (annualized) 1
Currency loss (annualized)
Res
t o
f A
fric
aCocoa 84 1,106,231 9.85% 9.16% 1.13% 1.61% 4.58%
Coffee 43 430,156 9.50% 8.67% 1.34% 2.00% 0.00%
Domestic Oriented
40 259,890 14.22% 10.00% 0.50% 6.75% 2.41%
Export Oriented 190 570,379 9.07% 6.99% 0.58% 5.13% 0.00%
East
Afr
ica
Cocoa 21 705,225 8.77% 7.91% 2.37% 0.00% 0.00%
Coffee 222 657,690 9.27% 7.68% 1.41% 6.65% 0.00%
Domestic Oriented
86 169,456 16.08% 10.18% 0.38% 7.25% 4.64%
Export Oriented 120 611,515 9.76% 8.37% 1.48% 3.82% 0.00%
Lati
n A
mer
ica/
Car
ibb
ean
Cocoa 164 602,613 8.00% 6.17% 0.86% 0.00% 0.00%
Coffee 1683 662,121 7.02% 6.16% 1.32% 4.40% 0.00%
Domestic Oriented
16 905,857 12.11% 7.41% 0.01% 0.00% 3.10%
Export Oriented 599 680,054 7.03% 5.53% 0.57% 1.09% 0.00%
Asi
a
Cocoa 11 525,909 9.01% 7.07% 0.77% 5.70% 0.00%
Coffee 78 834,463 9.29% 8.71% 1.21% 2.56% 0.02%
Domestic Oriented
10 433,184 14.65% 9.52% 0.03% 0.00% 5.50%
Export Oriented 194 917,331 8.44% 6.93% 0.35% 3.72% 0.00%
Global 3,556 664,743 9.07% 7.67% 0.80% 3.43% 0.38%
Appendix 4 : All CSAF lenders data by finance type and loan size (2/2)
1. Based on weighted average of individual loan yields by dollar duration
Product typeLoan Size Segment
# loansAverage loan
sizeHeadline
Interest Rate2
Interest yield p.a. (annualized,
gross) 1
Fee yield p.a. (annualized) 1
credit losses, net of recoveries (annualized) 1
Currency loss (annualized)
Ass
et f
inan
ce
equ
ipm
ent
$250k 176 137,174 9.67% 7.38% 0.33% 5.94% 0.55%
$250-500k 109 382,948 9.46% 8.16% 0.21% 3.42% 0.27%
$500k-1M 69 761,934 9.81% 6.76% 0.12% 3.75% 1.06%
>$1M 61 2,412,811 8.50% 6.38% 0.08% 1.77% 0.59%
Wo
rkin
g ca
pit
al
wit
h t
eno
r <
6
mo
nth
s
$250k 215 163,686 10.07% 9.45% 1.29% 5.65% 0.00%
$250-500k 320 381,976 9.73% 8.40% 1.14% 5.93% 0.00%
$500k-1M 253 786,761 8.93% 7.66% 0.94% 2.18% 0.00%
>$1M 122 1,618,841 8.44% 8.12% 0.91% 0.98% 0.00%
Wo
rkin
g ca
pit
al
wit
h t
eno
r <
12
m
on
ths
$250k 428 154,855 11.09% 10.70% 1.70% 3.48% 0.45%
$250-500k 464 392,859 9.64% 9.72% 1.42% 2.57% 0.00%
$500k-1M 435 798,262 9.29% 8.82% 1.36% 2.10% 0.01%
>$1M 301 1,809,318 8.45% 7.46% 1.15% 3.12% 0.00%
Wo
rkin
g ca
pit
al
wit
h t
eno
r >
12
m
on
ths
$250k 202 152,111 10.94% 8.35% 0.92% 7.94% 0.46%
$250-500k 174 394,286 9.85% 7.67% 1.05% 8.41% 0.14%
$500k-1M 122 773,497 9.43% 7.81% 1.00% 7.42% 0.10%
>$1M 110 1,940,240 8.60% 7.74% 1.01% 3.14% 1.13%
Global 3561 664,743 9.07% 7.67% 0.80% 3.43% 0.38%
Appendix 5 : Risk factor distribution
76
CSAF loan portfolio prevalence of risk factors
Number of loans, by number of risk factors
575
1,236
918
520
239
68
0
1,000
2,000
30 41 2 5
16% 35% 26% 15% 7% 2%% of
portfolio
Sub-Saharan Africa
New borrower
Loose value chain
Small loan
Long tenor
Risk Segments
Appendix 6: Glossary of financial performance terms
Annualized credit-losses yield. The total amount of credit losses as a proportion of the total dollar-duration of the portfolio. A credit
loss yield p.a. of 1% means that for every dollar that stays outstanding for a year, 1 cents will be lost due to credit losses.
Annualized fee income yield. The total amount of fee income received as a proportion of the total dollar-duration of the portfolio. A
fee income yield p.a. of 1% means that for every dollar that stays outstanding for a year, 1 cents will be received in fee income.
Annualized interest income yield. The total amount of interest income received as a proportion of the total dollar-duration of the
portfolio. An interest income yield p.a. of 1% means that for every dollar that stays outstanding for a year, 1 cents will be received in fee
income.
Annualized profitability yield. The total amount of profit received as a proportion of the total dollar-duration of the portfolio. A
profitability yield p.a. of 1% means that every dollar that stays outstanding for a year will make 1 cents of profit.
Credit losses. For a given portfolio, the total amount of money written-off plus expected credit losses. Expected write-offs are thus
accounted for as actual write off.
Dollar-duration. The product of the duration of the loan with the total amount disbursed, expressed in dollars. A dollar duration of $1 is
equivalent to a loan of $1 that is fully outstanding for exactly one year.
Duration. The average length of time that a given dollar of principal is outstanding. For example, a $1m loan being repaid in $500k
increments after 6 and 12 months has duration of 9 months.
Utilization. The average outstanding balance over the tenor of the loan. A 100% utilization would imply the full amount of the loan was
outstanding for the entirety of the loan tenor.
77
Appendix 7: Annualized yields for loans in Sub-Saharan Africa vs. rest of
the worldR
es
t o
f W
orl
dS
/Sa
ha
ran
Afr
ica
8.3%1.9%
-4.6%
-7.7%
6.5%
6.4%
3.2%
8.0%4.0%
0.7%
-2.4%
4.1%
3.3%
3.0%
Income net
of op. costs
Fee + interest Operating
costs
Recovery
costs +
write-offs
Income net of
bad debt costs
Risk-adjusted
impact cost of
funds*
Income net
of impact
cost of funds
78Note: (*) Impact cost of funds used is 3%
Source: CSAF lenders survey conducted between April – June, 2018 of 3,561 individual loan transactions
Currency loss=
Loan economics averages for all CSAF loans analyzed by region
Annualized yield (% per dollar p.a.)
Appendix 7: Annualized yields for small vs. large loan sizesL
oa
n s
ize
>$5
00
kL
oa
n s
ize
≤$500k
79Note: (*) Impact cost of funds used is 3%
Source: CSAF lenders survey conducted between April – June, 2018 of 3,561 individual loan transactions
Currency loss=
Loan economics averages for all CSAF loans analyzed by loan size segments
Annualized yield (% per dollar p.a.)
9.3%
-1.7%
-8.2%
-11.4%
3.2%
6.5%
11.1%
7.7%
5.3%
2.2%
-0.8%
2.4%
3.1%
3.0%
Fee + interest Income net
of op. costs
Operating
costs
Recovery
costs +
write-offs
Risk-adjusted
impact cost of
funds*
Income net of
bad debt costs
Income net
of impact
cost of funds
Appendix 7: Annualized yields for loans to new vs. existing borrowersE
xis
tin
g b
orr
ow
ers
Ne
w b
orr
ow
ers
7.0%2.0%
-3.6%
-6.8%
5.1%
3.2%
5.6%
8.6%4.3%
1.1%
-1.9%
4.3%
3.2%
Impact costs
of funds*
Income net
of op. costs
Fee + interest Operating
costs
3.0%
Recovery
costs +
write-offs
Income net of
bad debt costs
Income net
of impact
cost of funds
80Note: (*) Impact cost of funds used is 3%
Source: CSAF lenders survey conducted between April – June, 2018 of 3,561 individual loan transactions
Loan economics averages for all CSAF loans analyzed by borrower type
Annualized yield (% per dollar p.a.)
Currency loss=
Appendix 7: Annualized yields for loans in ‘loose’ vs. ‘tight’ value chainsC
off
ee
/Co
co
a
Loan economics averages for all CSAF loans analyzed by value chain group
Annualized yield (% per dollar p.a.)
Oth
er1
7.1%3.3%
-0.8%
-3.9%
3.8%
4.1%
3.1%
9.2% 3.9%
0.2%
-2.8%
5.3%
3.8%
2.9%
Income net of
bad debt costs
Fee + interest Operating
costs
Income net
of op. costs
Recovery
costs +
write-offs
Impact costs
of funds*
Income net
of impact
cost of funds
81Note: (*) Impact cost of funds used is 3%
Source: CSAF lenders survey conducted between April – June, 2018 of 3,561 individual loan transactions
Currency loss=
Appendix 7: Annualized yields for long-term vs. short-term loansTe
no
r ≤1
2 m
on
ths
Te
no
r >
12
mo
nth
s
7.0%3.4%
-1.1%
-4.3%
3.6%
4.4%
3.3%
9.5% 3.9%
0.6%
-2.2%
5.6%
3.3%
Recovery
costs +
write-offs
Fee + interest Impact costs
of funds*
Operating
costs
2.8%
Income net
of op. costs
Income net of
bad debt costs
Income net
of impact
cost of funds
82Note: (*) Impact cost of funds used is 3%
Source: CSAF lenders survey conducted between April – June, 2018 of 3,561 individual loan transactions
Loan economics averages for all CSAF loans analyzed by tenor segments
Annualized yield (% per dollar p.a.)
Currency loss=
top related